^{page 153 note 1} About ten minutes' walk to the North East o. Schimatari Junction, where the new railway line to Chalcis branches off from that which leads to Schimatari village and, ultimately, to Thebes.

^{page 154 note 1} So Mr. Papametrou's son informed me on the day I first saw it, April 24, 1905. His father has since made the same statement to Mr. R. C. Bosanquet.

^{page 154 note 2} A grant from the Craven Fund of the University of Oxford will make it possible for excavations to be carried out on the site during the spring of 1906 by Mr. A. C. Brown and other members of the British School at Athens, and plan and details may be reserved for the examination which will then be made of the whole neighbourhood. I must express my gratitude to Mr. J. L. Myres for the suggestion to which this grant is due. I am myself unfortunately unable to visit Greece again this spring.

^{page 154 note 3} The accompanying photograph and drawing are due to the kindness of Mr. R. C. Bosanquet, Director of the British School at Athens.

^{page 154 note 4} Scale, 1: 10, Letters ·06 m. high. Last stroke of N, ·045 to ·05 high.

^{page 154 note 5} I.G.S. (= I.G. vii.) 791 and 794.

^{page 155 note 1} So I.G.A. 509 (Roberts, i. 110) Κλεο[μέν]ης ἐποίησε τὠπ[έλ (λ)]ωνι on the steps of the Temple of Apollo at Syracuse.

^{page 155 note 2} So Mr. Bosanquet, and also Professor Ernest Gardner, who has kindly examined the photographs and drawing for me. He notices in particular the Ν and rather small ο.

^{page 155 note 3} Paus. ix. 22. 1. Nothing further seems to be known about this Temple. See B.C.H 1878, p. 339. There are inscriptions found at Tanagra relating to Artemis, e.g. I.G.S. (= I.G. vii.) 555. Αρτάμιδι Εἰλειθυίῃ and ib. 546 Αρτάμιδος. There was, however, a celebrated shrine of Artemis in the Tanagraean territory, at Aulis. Thus ib. 565 honours a woman ἱερατεύσασαν ᾿Αρτέμιδι Αὐλιδείᾳ and the representations of Artemis on Imperial Coins of Tanagra (J.H.S. viii. p. 9) clearly refer to that Temple. On the rim of a λέβης found at Tanagra we read Δεμοθερες ἱαρον Απολονος Καρυκεϝιο which Mr. Kouroniotes in Ἐφ. Ἀρχ. 1900, p. 107 suggests may refer to a cult of Apollo at Tanagra. He compares Ἐφ. Ἀρχ. 1896, p. 243 = B.C.H. 1896, pp. 242–3.

A comparison of Pausanias' two short ‘notebook’ statements in different chapters (22. 1 and 20. 1) as to the conjunction of Artemis and Leto with Apollo ‘at Tanagra,’ and their worship ‘at Delium in the territory at Tanagra’ gives an uneasy feeling that there may possibly here be a doublet. Curiously enough Imhoof-Blumer and P. Gardner in their Numismatic Commentary on Pausanias (J.H.S. viii. p. 11), when discussing the archaic figure of Apollo that appears on an Imperial Coin of Tanagra and arguing that it may be a copy of the statue at Delium, quote, not ix. 20. 1, but ix. 22. 1. Is that a slip, or a tacit anticipation of the suggestion here thrown out?

^{page 155 note 4} Stones were taken from Tanagra as far as the modern Oropo. See Milchhoeffer in Curtius and Kaupert's Attika, Text, Heft. ix. p. 25 and I.G.S. (= I.G. vii.) 510–512. The black stone, some blocks Of which lie round our chapel, is like that which was quarried near Tanagra and used there in great quantities. It was, however, common to the neighbourhood, and does not necessarily imply the bringing of stones from other buildings.

^{page 156 note 1} T. iv. 97. 2.

^{page 156 note 2} Frazer, Pausanias, v. p. 75.

^{page 157 note 1} Reisen und Forschungen, vol. ii. (1863) p. 46. He also saw remains of a jetty or pier, which is good evidence for a port, but throws no light on the Temple. See below, p. 159, n. 5.

^{page 157 note 2} So Leake, , Northern Greece, ii. p. 453Google Scholar. Bursian, , Geog. i. pp. 218–9.Google Scholar Frazer, loc. cit. and all modern Histories and Guide Books. Ross, Wanderungen, p. 105 argued for Dramesi, a village on the coast nearer to Aulis, and Vischer, Erinnerungen, p. 678 doubted between the two, but their grounds were insufficient.

^{page 157 note 3} ix. 20. 1.

^{page 157 note 4} ix. p. 403. Scylax Periplus 59 should be added. For a discussion as to its value, see Tozer, Hist. Ancient Geog. p. 119.

^{page 157 note 5} vi. 118.

^{page 157 note 6} xxxv. 51.

^{page 157 note 7} See Appendix A.

^{page 157 note 8} x. 28. 6. ἀπέδωκεν Ταναγραίοις ἐς Δήλιον He would naturally assume that Herodotus had made a slip here, from his knowledge of the political situation as it existed in his own day. He would also remember the monument at Athens (i. 29. 13) and the narrative of Thucydides which showed that Delium belonged to Athens in the Peloponnesian War (iii. 6. 1, ix. 6. 3, cp. vi. 19. 5). We cannot assume, however, that there is no truth in Herodotus' statement. From coins (B.M.Cat. Central Greece, Introduction, xxxviii.) we learn one phase of the internal politics of Boeotia from B.C. 480 to 456, the aggrandisement of Tanagra at the expense of Thebes. Mr. Percy Ure has suggested to me that here in the δἰ ἐτέων εἴκοσι of Herodotus (vi. 118), we may have an echo of a counter move on the part of Thebes. The action of Sparta in 457 would be in support of this policy, and perhaps consequent on the hostility it had aroused.

^{page 158 note 1} Paus. ix. 19. 5. I hope to discuss this portion of his route later in a separate article.

^{page 158 note 2} So Vogel, A. in Philologus, 1882, p. 516Google Scholar; Niese, in Hermes, 1878, p. 43Google Scholar; Tozer, Selections from Strabo, p. 18. A plea for a visit to Elis is urged by Schroeter, De Strabonis Itineribus, 1874, but Frazer, Paus. i. Introduction xci is right when he sums up: ‘It is generally recognized that Strabo visited very few parts of Greece, perhaps none but Corinth.’

^{page 158 note 3} So Ulrichs, op. cit. p. 47, n. 1.

^{page 158 note 4} See Appendix B.

^{page 159 note 1} iv. 76. 4 and 90. 1. This repeated insistence on the definition attracted the notice of Classen, who proposed, without MS. authority, to excise it in the second passage. He remarks that if Thucydides had wished to give the ‘superfluous’ explanation, he would certainly have attached it to the general expression of direction, ἐπὶ τὸ Δήλιον which immediately precedes it. So he probably would, if there had not been some special reason, such as has been suggested.

^{page 159 note 2} See Appendix C.

^{page 159 note 3} T. iv. 90. 2.

^{page 159 note 4} As the whole district of Artemisium was called after its Temple (Hdt. vii. 176, quoted above). So the coast by Delium itself was sometimes called Tanagra. Diitrephes (T. vii. 29. 2), before passing the Euripus, landed his Thracians ἐς τὴν Τάναγραν For the Temple property, notice the sacred ‘lucus’ in Livy, xxxv. 51.

^{page 159 note 5} For the general tendency, see T. i. 7. 7. It affected cities already crystallized, such as Athens, but naturally took a different form. The special cause in this case may have been the expansion of Tanagra and its need of a further outlet to the sea. The Valley of Dilisi would serve for such a purpose. It has been already noticed that the only important remains found there by Ulrichs were those of a jetty or pier, see above p. 157, n. 1.

^{page 160 note 1} It was still a centre of worship. T. iv. 97. 2, 98. 2.

^{page 160 note 2} This would affect Herodotus as well as Thucydides.

^{page 160 note 3} iv. 76. 3. We must remember that it is not only improbable but almost impossible that Thucydides was at Delium. He must have already started for Thasos (iv. 104. 3).

^{page 160 note 4} iv. 76. 4. πρὸς Εὔβοιαν τετραμμένον See ii. 55. 2 and especially vii. 58. 1. τῆς Σικελίασ τὸ πρὸς Λιβύην μέρος τετραμμένον νεμόμενοι

^{page 160 note 5} See Appendix D.

^{page 160 note 6} 216 A. That Athenaeus has got Thucydides' own narrative in his mind, if not before him, is shown by his using his direct words, especially where Pagondas sends round δύο τέλη ἐκ τοῦ ἀφανοῦς. Below he has a variant οἷ ἱππεῖς οἹ τε αὐτῶν (ὅτι ἑαυτῶν is clearly a MS error) καὶ Λοκρῶν for Thucydides' … καὶ οἱ Λοκροί. It is improbable that Herodicus δ῾ Κρατήτειος, whom he has just been using, had occasion to quote Thucydides at such length.

^{page 160 note 7} xii. 70. 4. For Ephorus as source of Diodorus, see Busolt, , Gr. G. iii. 2, pp. 707–710Google Scholar; Beloch, i. p. 19; Volquardsen, Die Quellen Diod. 1868; Collmann, de Diod. fontibus, 1869. For some modifications of the extreme view, see Brōcker, Untersuchungen über Diod. 1879.

^{page 161 note 1} For Ephorus' interest in Boeotia, see Diod. xi. 82, Strabo ix. 400, Plut. περὶ Ἀδολεσχίας 22. There are four points in Diodorus' narrative which certainly come from a Boeotian source: (1) § 1. The chosen 300, ἡνίοχοι καὶ παραβάται καλούμενοι. (2) § 3. The Thebans διαφέρουτες ταῖς τῶν σωμάτων ῥώμαις. (3) § 5. The great stoa in the ἀγορά of Thebes decorated with the price of the spoils. (4) § 5. The foundation of the πανήγυρις τῶν Δηλὶων cp. I.G.S. (= I.G. vii.) 20.

There is no reason to doubt the value of these points, but the charge of the Athenian horse at the beginning of the engagement, whether from the same source, or a piece of stock military description due to Ephorus or Diodorus seems inconsistent with Thucydides and improbable in itself. Should one trace a Boeotian version in Plut. De Gen. Soc. 11., where Pyrilampes, while a prisoner in Boeotia (cp. Thuc. v. 35), hears the story of Socrates' escape?

^{page 161 note 2} See Appendix E.

^{page 162 note 1} Another partial explanation is conceivable, that the coast-line has changed since classical times. The Geological Maps published by Dr. Alfred Philippson and DrSkuphos, T. G.. (Zeitschrift der Gesellschaft für Erdkunde zu Berlin, vol. xxv. 1890, p. 406 and xxix. 1894, p. 409)Google Scholar adapted from the same earlier material with independent observations, mark the coast and immediate hinterland from Vathy to the other side of Dilisi as alluvial. Though there does not seem to be enough river action in the district to warrant suggesting that as a cause, the shallow soundings on the Admiralty Charts (1881) may point to the fact that the current of the Euripus drives deposits. into the slack water of the Bay. The whole neighbourhood too has been in a seismic zone from the days of Thucydides (iii. 87. 89) to our own. In the earthquake of April, 1894, the coast ο Lokris was covered by the sea for an extent of 6 miles long by 300 yards broad (Skuphos, op. cit. p. 438), and the half island Gaïdarion was permanently made into an island. The reverse process, the rising of land, is equally possible.

The fact that Geology appeals to History for the dating of other changes on the north-east coast-line of Greece makes this a reasonable line of enquiry. In remarking on the enormous alluvial deposits of the Spercheius, Dr. Philippson (op. cit. p. 386) can only fix the geological date by the historical fact that in classical times Thermopylae ‘must have been’ on the sea.

^{page 162 note 2} Quum … alii ad spectaculum templi lucique versi, alii in litore inermes vagarentur, magna pars per agros … dilapsa esset … There is nothing to necessitate the Temple being on the shore.

^{page 163 note 1} T. iii. 91. 3–4. That Skala Oropou and not Oropou is the ancient Fifth Century Oropus is (pace Baedeker, ed. 5, 1905, p. 167) clear from Curtius-Kaupert, Text, Heft ix. pp. 23. 4; Frazer, , Paus. ii. pp. 465–6.Google Scholar

^{page 163 note 2} T. iv. 76. 2. So Grote, vi. p. 160, though without getting to close quarters with the topography, assumes that Delium was ‘strongly situated, overhanging the sea.’ Later writers, such as Frazer and Busolt, have ignored this point.

^{page 163 note 3} iii. 2, p. 1114 with note 3.

^{page 164 note 1} T. iv. 42. 2. For the Topography, see Leake, , Morea, vol. iii. p. 309Google Scholar, Curtius, Pelop. 548 ff. Arnold, , Thuc. vol. 2, p. 443 and Map at p. 54.Google Scholar

^{page 164 note 2} T. iii. 91. 3.

^{page 164 note 3} T. iv. 76. 1. There would of course be small squadrons in different parts of the empire, such as Thucydides' seven ships at Thasos (ib. 104. 4).

^{page 164 note 4} T. iv. 77. 1, 90. 1.

^{page 164 note 5} T. iv. 101. 1. It may be asked why they left a garrison in so insecure a position. We can only suppose that they hoped vaguely, as at Plataea, that the Boeotians would fail in a siege.

^{page 164 note 6} T. iv. 100. 5. As we do not know how large the garrison was, we can infer nothing further rom Thucydides' one figure, the 200 who were captured.

^{page 164 note 7} Op. cit. p. 49. Another road leads along the sea-shore to the ancient Oropus. But it is certain for more reasons than one that the Athenians were making for Attica, and not for Oropus, and that the battle was not fought on the sea. Many of the fugitives for instance—Socrates and the rest—made for Parnes, and they would not have done so if the walls of Oropus had been on the direct road in front of them. Unless, too, with Grote (vi. p. 161), the battle is placed 25 or 30 stades from Dilisi, at the west extremity of the plain of Oropus, the ground is not suited for the battle. If the Temple is to be placed in or close to the valley of Dilisi, these arguments also apply to a track that runs between the inland and the coast roads and issues at Sykamino, at the head or south-west end of the plain of Oropus. Such a route to the passes of Parnes would be at once much longer and rougher, and would only be the natural one if the Temple were placed on the coast at least a kilometre east of the valley of Dilisi. The heights here are worth exploring. The chief difficulties would be that the position is not a good one or commanding the Tanagraean plain, and the ground at a reasonable distance is over rugged for the battle; the chief advantage, that it seems more easy to suppose that within ten stades the army would pass into the territory of Oropus. We have no direct evidence as to where the boundary lay (see Curt.-Kaup. Att. Text, Heft ix. p. 16). That Oropus had considerable territory to the west of it is suggested by T. ii. 23, 3, where the Spartans, on leaving Attica for Boeotia, ravage territory ἣν νέμονται Ὠρώποιο,παοιόντες Ὠρωπόν, passing Oropus. See too Curt.-Kaup. op. cit. p. 30 for the Boeotian boundaries southeast of Tanagra. It is difficult to judge how much weight we should attach to the modern line of boundary between the Eparchies of Attica and Thebes, as given in the French Carte de la Grèce, 1852. South of the Asopus it gives ground west of our battlefield to Attica, striking it slightly west of the mediaeval tower on the Hill of Staniates. It then turns one kilometre due east, and after a run of two kilometres almost due north, edges east again, and reaches the sea about one kilometre east of Dilisi. If the exact course of this boundary rests on continuous tradition, it seems slightly against our suggested battleground, which it passes about one kilometre to the east. This line of argument, however, is not supported by the consideration of other facts; such, for instance, as that Eleusis is placed in the Eparchy of Megara, and not in that of Attica.

It must be remembered, further, that the ‘boundaries of Oropia’ in which the Boeotians saw the Athenian force (T. iv. 91) would not necessarily be the limited ground on which the battle was finally fought, but the considerable space covered by a large army on the march, with its ψιλοί (ib. 90. 4, 94. 1) spread out in front of it. The casualties, too, mostly occurred during the flight (ib. 96. 7, 101. 2) and the great majority of the dead bodies would certainly be in Oropia, even were the battle itself slightly west of it.

^{page 165 note 1} Striking the main road about 2½ kilometres east of Staniates Tower. Neither the inland road from Dilisi, nor the present main road, as given in Curtius-Kaupert, Attika Kart., Sect. Oropos, No. 9, is marked in the French Map of 1852. In the former case this is probably due to its insignificance, in the latter to its having been made since 1852. The main road, as marked in the French Map, crossed the Asopus nearly three kilometres further east than the present one.

^{page 165 note 2} T. iv. 93, 1–2.

^{page 166 note 1} Aristeides Dedouses, who has an intimate knowledge of Tomb excavation in the Tanagra district during the last decades, informed me that lines of ancient tombs have been found along its course. I myself picked up fragments of lustrous black glaze pottery at various points along it. I may mention that Dedouses had never seen the Apollo Inscription till I showed him it on my second visit on April 26.

^{page 166 note 2} T. iv. 93. 3.

^{page 166 note 3} T. iv. 93. 3, ὑπερεφάνησαν; 96. 5, ὑπερφανέντων.

^{page 167 note 1} The Athenians had a front of 875 hoplites (T. iv. 94. 1). The Boeotian hoplite line was shorter because of the depth of the Theban contingent (ib. 93. 4), but probably some of their lightarmed filled the gap, or possibly the ground narrowed on their side. The surrounding of the Thespians on the Boeotian left (ib. 96. 3) may point to the fact that the Athenian hoplite line overlapped on the right. We may be sure that the Thebans, according to the general tendency to guard the unshielded side, would edge to the limit of the rough ground on their right. Delbrück (Kriegskunst, pp. 96. 7) remarks that the fact that the Boeotian ψιλοί were stopped by the ῥύακες shows they had not much warlike energy. But the fact was that there were few or no Athenians ψιλοί opposite them (ib. T. iv. 94. 1) and such troops had not yet learnt they could interfere in a hoplite engagement. Kromayer (Antike Schlachtfelder, p. 321) is probably right in claiming a front of 1 metre for each hoplite as against Delbriick's ·44, though all the reasons he urges for his conclusion need not be accepted.

^{page 167 note 2} T. iv. 96. 2.

^{page 167 note 3} T. iv. 93. 4, 94. 1.

^{page 167 note 4} T. iv. 96. 5.

^{page 167 note 5} T. iv. 89. 1.

^{page 167 note 6} op. cit. p. 49.

^{page 168 note 1} Paus. v. p. 76.

^{page 168 note 2} III. 2. 1147, n. 1.

^{page 168 note 3} The hill on which excavations conducted three years ago laid bare some ancient foundations, probably of houses. Mr. Bosanquet has pointed out that the name clearly refers to the stone quarried there, as Kiséri, ancient κίσηπις, properly means pumice stone, and is then apparently used. in Boeotia for a soft light free-stone.

^{page 168 note 4} Only with the proviso that no deep gullies are demanded by the situation. The arable ground, too, may well have changed.

^{page 168 note 5} There is a rough forest-track further west which is a possible but not probable route from Dilisi.